Hyper-Locating Places-Of-Interest in Buildings

1. A computer-implemented method of hyper-locating places-of-interest within a building structure, the method including: outlining a contour of the building structure based on geospatial coordinates of perimeter points of the building structure; receiving preliminary geospatial coordinates of a plurality of places-of-interest and qualifying, as contained, the places-of-interest that are within the contour; fitting one or more regression lines to the contained places-of-interest; generating adjusted places-of-interest by projecting the contained places-of-interest onto a nearest regression line; feeding the adjusted places-of-interest as input seeds to a Voronoi block generator to partition the building contour into blocks, each block associated with a block size; and storing respective centroids of the blocks as hyper-located geospatial coordinates of the contained places-of-interest.

2. The computer-implemented method of claim 1 , wherein the blocks are Voronoi cells.

3. The computer-implemented method of claim 1 , further including: assigning each of the adjusted places-of-interest to a corresponding block.

4. The computer-implemented method of claim 1 , further including: estimating principal axes of the building structure determined from principal component analysis the geospatial coordinates of the perimeter points and/or the contained places-of-interest; and dividing the contour into pieces and calculating a regression line for each of the pieces along the principal axes.

5. The computer-implemented method of claim 1 , further including using spatial indices for qualifying, as contained, the places-of-interest that are within the contour, wherein the qualification is parameterized by a radius around the geospatial coordinates of the perimeter points, and wherein the radius is two meters.

6. The computer-implemented method of claim 5 , further including qualifying at least some of the contained places-of-interest as select places-of-interests belonging to recognized brands, and adjusting block sizes for the select places-of-interests to known areas of location units of the recognized brands.

7. The computer-implemented method of claim 1 , wherein each of the blocks is the same block size.

8. The computer-implemented method of claim 7 , wherein the block sizes have a minimum area of hundred square meters.

9. The computer-implemented method of claim 1 , wherein the outlining further includes sampling geospatial coordinates of additional perimeter points along distribution of the geospatial coordinates of the perimeter points with a step size, and wherein the step size of the sampling is twenty meters.

10. The computer-implemented method of claim 1 , wherein the geospatial coordinates of a particular one of the centroids of a particular one of the blocks are determined based on a mean of geospatial coordinates of vertices of the particular one of the blocks.

11. The computer-implemented method of claim 1 , further including selecting building structures for partitioning based on at least one of a minimum area of the building structures, geographic location of the building structures, elevation of the building structures, and a minimum number of the contained places-of-interest.

12. The computer-implemented method of claim 1 , further including: receiving, from different data sources, multiple sets of geospatial coordinates of perimeter points for a particular building structure; and selecting at least one set from the multiple sets for outlining the contour of the particular building structure based on trustworthiness of the sources.

13. The computer-implemented method of claim 1 , further including performing supplementary partitioning for certain ones of the contained places-of-interest.

14. The computer-implemented method of claim 1 , wherein a count of the blocks is determined based on a count of the contained places-of-interest.

15. The computer-implemented method of claim 1 , further including buffering outlines of the contour inwards to ensure that the regression lines are within the contour.

16. The computer-implemented method of claim 1 , further including correlating the hyper-located geospatial coordinates of the contained places-of-interest with geospatial coordinates of visitor locations to detect visitor visits to the contained places-of-interest, to determine whether a mobile device was inside or outside of the building structure at a given timestamp, to determine whether a mobile device was inside or outside a particular contained places-of-interest at a given timestamp, to determine a duration for which a mobile device dwelled at a particular contained places-of-interest, and/or to determine that a mobile device visited a first building structure prior to visiting a second building structure.

17. A computer-implemented method of hyper-locating places-of-interest within a building structure, the method including: outlining a contour of the building structure based on geospatial coordinates of perimeter points of the building structure; receiving preliminary geospatial coordinates of a plurality of places-of-interest and qualifying, as contained, the places-of-interest that are within the contour; sampling random points within the contour and iteratively projecting line segments through some of the random points to decompose the contour into disjoint maximum-area rectangles; determining a regression line for each of the rectangles; generating adjusted places-of-interest by projecting the contained places-of-interest onto a corresponding regression line; feeding the adjusted places-of-interest as input seeds to a Voronoi generator and partitioning the building contour into blocks; and storing respective centroids of the blocks as hyper-located geospatial coordinates of the contained places-of-interest.

18. The computer-implemented method of claim 17 , further including: determining whether a portion of the contour outlined by the iterative projection of the line segments is a rectangle based on fitting a bounding rectangle over the portion and calculating an area overlap between the bounding rectangle and the portion; and classifying the portion as the rectangle when the area overlap is equal to or above a threshold, and wherein the threshold is ninety-five percent.

19. The computer-implemented method of claim 18 , further including progressively increasing contour area enclosed by the portion until one of the disjoint maximum-area rectangles is outlined by the iterative projection.

20. The computer-implemented method of claim 19 , further including terminating the iterative projection when the contour area cumulatively enclosed by the disjoint maximum-area rectangles exceeds a threshold, and wherein the threshold is ninety percent.